method · first seen 14 Oct 2026
reinforcement learning from measured performance for code optimization
Training a code-generating model with a reward combining correctness and measured runtime speedup of generated kernels.
Mentions per week
PapersCodeDiscussion
Earliest sources
- How to Train Your Little Dragon: Teaching a Small Language Model to Construct and Refine GPU Optimization ScriptsPapers · 14 Oct 2026 · Annabelle Sujun Tang, Wei Tang, Songyao Jiang, Tianhui Li, Shengzhe Xu, Jiin Woo…
Latest sources
- How to Train Your Little Dragon: Teaching a Small Language Model to Construct and Refine GPU Optimization ScriptsPapers · 14 Oct 2026 · Annabelle Sujun Tang, Wei Tang, Songyao Jiang, Tianhui Li, Shengzhe Xu, Jiin Woo…